Caching introduces a consistency problem: cached data can become stale while the server contains newer data.

Applications can address this through explicit cache updates, refetching, time-based policies, subscriptions, polling, cache invalidation, or server-driven events.

There is no universally correct caching policy. Highly dynamic data often requires more aggressive revalidation, while relatively stable data can remain cached for longer.